Direkt zum Inhalt springen
login.png Login join.png Register    |
de | en
Technical University of Munich

Technical University of Munich

Sitemap > Bulletin Board > Diplomarbeiten, Bachelor- und Masterarbeiten > Last-mile crowdsourced parcel delivery logistics (Master Thesis)
up   Back to  News Board    previous   Browse in News  next    

Last-mile crowdsourced parcel delivery logistics (Master Thesis)

01.10.2021, Diplomarbeiten, Bachelor- und Masterarbeiten

A master thesis topic offered by the Professorship of Operations and Supply-Chain Management (Prof. Schiffer), TUM School of Management. We are looking for students of the TUM-BWL (with a major in Supply Chain Management), Informatics, Engineering, or similar study programs, with interest in topics of Operations Research and AI.

To cope with increased e-commerce sales volumes and customer expectations with regards to same-day or on-demand delivery, a number of businesses rely on crowdsourced couriers to perform last-mile logistics. Within crowdsourced delivery, so-called on-the-way delivery services are emerging. Here the company in charge of the delivery, typically a logistics service provider (LSP), matches delivery requests to already existing routes of private individuals, e.g., when they are driving home from work (see, e.g., Roadie). These individuals receive a small compensation in return for the delivery and the detour from their original route. In contrast to more established crowdsourced delivery models (e.g., Amazon Flex or Postmates) no new routes are created for making deliveries. This helps in reducing increasing traffic from delivery logistics and in lowering the LSP’s operational costs.

However, operating a fleet of crowdsourced drivers, in the literature also referred to as occasional drivers (OD), comes with significant challenges. While the LSP tries to minimize its operational costs, ODs try to maximize their earnings. Existing research focuses mainly on the operational costs of the LSP and formulates the problem as a vehicle routing problem with occasional drivers (VRPOD). The behavior of occasional drivers (OD) and their objectives are typically neglected. The goal of this work is therefore to consider ODs’ objective and constraints in the analysis, e.g., by incorporating uncertainty in drivers’ availability.

Aims and scope of the thesis

It is the subject of this thesis to model a pick-up and delivery problem in which the LSP operates a mixed fleet of ODs and dedicated drivers (i.e., full-time employees). It is assumed that the availability of drivers and requests as well as their origins and destinations are known beforehand. The compensation of an OD consists of a fixed and a detour-dependent term. While the LSP tries to minimize its operational costs, the OD wants to maximize the ratio between compensation and detour made. This can be formulated as a bi-level optimization problem, where the LSP’s optimization problem is the first level problem and the OD’s optimization problem is the second. The bi-level optimization problem can then be reformulated as a single-level problem with the help of the KARUSH-KUHN-TUCKER (KKT) conditions and solved with a, e.g., metaheuristic. To summarize, the work consists of the following research tasks:

  • Review of literature concerning the VRPOD
  • Formally define the bi-level optimization problem
  • Implement a solution method, e.g., metaheuristic, to solve instances of reasonable sizes
  • Design and conduct experiments
  • Compare benefits of such an approach with literature


This thesis targets students of the TUM-BWL (with a major in Supply Chain Management), Informatics, Engineering, or similar study programs. Knowledge of mathematical programming, optimization, and a general-purpose programming language (e.g. C++, Java, Python) is required. Prior participation in one of the seminars offered by the chair (i.e. Modeling Future Mobility Systems, Advanced Seminar) is recommended. The thesis should be written in English.

Related Research

  • Archetti C, Savelsbergh M, Speranza G (2016) The vehicle routing problem with occasional drivers. European Journal of Operational Research 254(2):472–480.
  • Arslan A, Zuidwijk R (2016) Crowdsourced delivery—a dynamic pickup and delivery problem with ad hoc drivers. Transportation Science 53(1):472–480.
  • Allende G.B., Still G. (2013) Solving bilevel programs with the KKT-approach. Mathematical Programming 138:309–332.

Begin: as soon as possible

Application: To apply, please follow the process described on our website

Kontakt: julius.luy@tum.de

More Information


Todays events

no events today.

Calendar of events